Attention mechanism has been a key component in Recurrent Neural Networks (RNNs) based sequence to sequence learning framework, which has been adopted in many text understanding tasks, such as neural machine translation and abstractive summarization. In these tasks, the attention mechanism models how important each part of the source sentence is to generate a target side word. To compute such importance scores, the attention mechanism summarizes the source side information in the encoder RNN hidden states (i.e., h_t), and then builds a context vector for a target side word upon a subsequence representation of the source sentence, since h_t actually summarizes the information of the subsequence containing the first t-th words in the source sentence. We in this paper, show that an additional attention mechanism called word attention, that builds itself upon word level representations, significantly enhances the performance of sequence to sequence learning. Our word attention can enrich the source side contextual representation by directly promoting the clean word level information in each step. Furthermore, we propose to use contextual gates to dynamically combine the subsequence level and word level contextual information. Experimental results on abstractive summarization and neural machine translation show that word attention significantly improve over strong baselines.

Neural machine translation (NMT) heavily relies on parallel bilingual data for training. Since large-scale, high-quality parallel corpora are usually costly to collect, it is appealing to exploit monolingual corpora to improve NMT. Inspired by the law of total probability, which connects the probability of a given target-side monolingual sentence to the conditional probability of translating from a source sentence to the target one, we propose to explicitly exploit this connection to learn from and regularize the training of NMT models using monolingual data. The key technical challenge of this approach is that there are exponentially many source sentences for a target monolingual sentence while computing the sum of the conditional probability given each possible source sentence. We address this challenge by leveraging the dual translation model (target-to-source translation) to sample several mostly likely source-side sentences and avoid enumerating all possible candidate source sentences. That is, we transfer the knowledge contained in the dual model to boost the training of the primal model (source-to-target translation), and we call such an approach dual transfer learning. Experiment results on English-French and German-English tasks demonstrate that dual transfer learning achieves significant improvement over several strong baselines and obtains new state-of-the-art results.

A notable challenge in cross-lingual wikification is the problem of retrieving English Wikipedia title candidates given a non-English mention, a step that requires translating names written in a foreign language into English. Creating training data for name translation requires significant amount of human efforts. In order to cover as many languages as possible, we propose a probabilistic model that leverages indirect supervision signals in a knowledge base. More specifically, the model learns name translation from title pairs obtained from the inter-language links in Wikipedia. The model jointly considers word alignment and word transliteration. Comparing to 6 other approaches on 9 languages, we show that the proposed model outperforms others not only on the transliteration metric, but also on the ability to generate target English titles for a cross-lingual wikifier. Consequently, as we show, it improves the end-to-end performance of a cross-lingual wikifier on the TAC 2016 EDL dataset.

Partially inspired by successful applications of variational recurrent neural networks, we propose a novel variational recurrent neural machine translation (VRNMT) model in this paper. Different from the variational NMT, VRNMT introduces a series of latent random variables to model the translation procedure of a sentence in a generative way, instead of a single latent variable. Specifically, the latent random variables are included into the hidden states of the NMT decoder with elements from the variational autoencoder. In this way, these variables are recurrently generated, which enables them to further capture strong and complex dependencies among the output translations at different timesteps. In order to deal with the challenges in performing efficient posterior inference and large-scale training during the incorporation of latent variables, we build a neural posterior approximator, and equip it with a reparameterization technique to estimate the variational lower bound. Experiments on Chinese-English and English-German translation tasks demonstrate that the proposed model achieves significant improvements over both the conventional and variational NMT models.

We describe a technique for structured prediction, based on canonical correlation analysis. Our learning algorithm finds two projections for the input and the output spaces that aim at projecting a given input and its correct output into points close to each other. We demonstrate our technique on a language-vision problem, namely the problem of giving a textual description to an "abstract scene".

However, the sentences under analysis may experiments, we find that both joint methods produce tree not always be grammatically correct. When a dependency fragment sets that are more similar to those produced by the parser nonetheless produces fully connected, syntactically oracle method than the previous pipeline method; moreover, well-formed trees for these sentences, the trees may be inappropriate the seq2seq method's pruning decision has a significantly and lead to errors. In fact, researchers have raised higher accuracy. In terms of downstream applications, we valid questions about the merit of annotating dependency show that dependency arc pruning is helpful for two applications: trees for ungrammatical sentences (Ragheb and Dickinson sentential grammaticality judgment and semantic role 2012; Cahill 2015). On the other hand, previous work has labeling.

Previous neural machine translation models used some heuristic search algorithms (e.g., beam search) in order to avoid solving the maximum a posteriori problem over translation sentences at test phase. In this paper, we propose the \textit{Gumbel-Greedy Decoding} which trains a generative network to predict translation under a trained model. We solve such a problem using the Gumbel-Softmax reparameterization, which makes our generative network differentiable and trainable through standard stochastic gradient methods. We empirically demonstrate that our proposed model is effective for generating sequences of discrete words.

While end-to-end neural machine translation (NMT) has achieved notable success in the past years in translating a handful of resource-rich language pairs, it still suffers from the data scarcity problem for low-resource language pairs and domains. To tackle this problem, we propose an interactive multimodal framework for zero-resource neural machine translation. Instead of being passively exposed to large amounts of parallel corpora, our learners (implemented as encoder-decoder architecture) engage in cooperative image description games, and thus develop their own image captioning or neural machine translation model from the need to communicate in order to succeed at the game. Experimental results on the IAPR-TC12 and Multi30K datasets show that the proposed learning mechanism significantly improves over the state-of-the-art methods.

Attention-based Encoder-Decoder has the effective architecture for neural machine translation (NMT), which typically relies on recurrent neural networks (RNN) to build the blocks that will be lately called by attentive reader during the decoding process. This design of encoder yields relatively uniform composition on source sentence, despite the gating mechanism employed in encoding RNN. On the other hand, we often hope the decoder to take pieces of source sentence at varying levels suiting its own linguistic structure: for example, we may want to take the entity name in its raw form while taking an idiom as a perfectly composed unit. Motivated by this demand, we propose Multi-channel Encoder (MCE), which enhances encoding components with different levels of composition. More specifically, in addition to the hidden state of encoding RNN, MCE takes 1) the original word embedding for raw encoding with no composition, and 2) a particular design of external memory in Neural Turing Machine NTM) for more complex composition, while all three encoding strategies are properly blended during decoding. Empirical study on Chinese-English translation shows that our model can improve by 6.52 BLEU points upon a strong open source NMT system: DL4MT1. On the WMT14 English-French task, our single shallow system achieves BLEU=38.8, comparable with the state-of-the-art deep models.

Pronouns are frequently omitted in pro-drop languages, such as Chinese, generally leading to significant challenges with respect to the production of complete translations. To date, very little attention has been paid to the dropped pronoun (DP) problem within neural machine translation (NMT). In this work, we propose a novel reconstruction-based approach to alleviating DP translation problems for NMT models. Firstly, DPs within all source sentences are automatically annotated with parallel information extracted from the bilingual training corpus. Next, the annotated source sentence is reconstructed from hidden representations in the NMT model. With auxiliary training objectives, in the terms of reconstruction scores, the parameters associated with the NMT model are guided to produce enhanced hidden representations that are encouraged as much as possible to embed annotated DP information. Experimental results on both Chinese-English and Japanese-English dialogue translation tasks show that the proposed approach significantly and consistently improves translation performance over a strong NMT baseline, which is directly built on the training data annotated with DPs.